General Instrument Care of the Viscotek TDAmax GPC/SEC system

Introduction

The Viscotek GPC/SEC systems (Fig. 1) combine the separating power of chromatography with the analytical power of light scattering detection and viscometry for size and structural characterization of molecules in solution.

Viscotek GPC/SEC systems are used to analyze a wide range of molecules, ranging from polymers to proteins. This versatility means that the systems can be run in many different conditions and solvent setups.

This technical note details information about maintaining Viscotek GPC/SEC systems and aims to help you:

• Maximize your system's reliability.
• Prevent damage to the instrument.
• Reduce unnecessary downtime.
• Recognize rising issues.
• Minimize avoidable service visits.

Organic Systems

Systems running in organic solvents are generally simple to maintain. The main danger of damage to the instrument occurs during changeover between two organic solvents. This must be performed according to the protocol outlined in Section 4 "Solvent compatibility and changeover" which can be found on pages 5 - 7 of this technical note.

Our recommended guidelines for Viscotek GPC/SEC systems running in organic solvents are set out in Table 1.

Aqueous Systems

The nature of running systems in aqueous buffers means that they are at risk of microbial contamination. If a microbial culture is allowed to grow in a system, it can disrupt the chromatography and can also damage the detectors. Whilst minor contamination can be eliminated by flushing the system with cleaning agents, major contamination will require a service visit.

Recirculating aqueous buffers or stopping the flow without flushing the system with a preservative buffer is not recommended. This action will significantly increase the likelihood of microbial contamination and quickly render the instrument unusable.

Table 2 sets out our recommended guidelines for Viscotek GPC/SEC systems running in aqueous buffers.

1. Solvent and sample preparation

Solvent preparation

Careful preparation of organic and aqueous solvents can significantly reduce the risk of particulate and microbial contamination.

We recommend that only HPLC grade solvents are used in organic systems. For aqueous systems, HPLC grade bottled water or ultra-pure water from a purification system must be used. It is essential that both organic and aqueous solvents are filtered through a 0.2 µm PTFE or nylon filter.

For aqueous running solvents, we recommend:

• The use of additives such as sodium azide (0.02% w/v), alcohols (methanol or ethanol, 10% v/v) or acetone and acetonitrile. However, these should only be used if they have been shown to be compatible with the separation method and/or analyte.
• That glassware is cleaned with detergent and rinsed with neat alcohol, and left to dry before use.
• That aqueous solvents are kept away from direct sunlight and stored at 4oC, if not immediately used.

Sample preparation

The most important point to consider with respect to sample preparation is that GPC/SEC is a solution-based technique. Consequently, good chromatography relies on the sample being fully dissolved. Samples should therefore be given sufficient time to fully dissolve.

Sample solutions must be clear, not cloudy or turbid. We recommend that 0.45 µm or 0.2 µm filters are used prior to sample injection. Typical filters are PTFE (Teflon) for samples in organic solvents and nylon or cellulose for samples in aqueous buffers.

2. Routine maintenance

From time to time, routine maintenance will be required in order to keep the system in the best condition. This should be planned to avoid analysis delays and sufficient time should be set aside.

Column exchange

Columns should be used and run according to the manufacturer's instructions. To exchange columns, the flow rate should be reduced to 0.1 ml/min or stopped. The columns can then be safely removed from the flow path.

In the TDA, the columns are kept inside the column oven. To access the columns, the DC power should be turned off, the screws on the front of the system loosened and the detector module carefully removed from the chassis.

If new columns require the solvent in them to be changed, the columns should be purged to waste for at least two column volumes before connecting them to the detector. Follow the column manufacturer's instructions regarding the solvent change. It is likely that some material will elute from the columns due to the new solvent. This should always be prevented from entering the detector as it may be a source of contamination.

Figure 2: Diagram of the components of the post-column filter.

The new columns are inserted into the flow path carefully to avoid allowing any air into the system. It is worth watching the connections for a couple of minutes to check for leaks. The detector module is then replaced, the screws tightened, and the DC power switched on. When the flow is accelerated, care should be taken to make sure the detectors are stable and do not exceed the measurement range. If any air has been introduced, the detectors signals will be moving randomly. They should be purged and allowed to return to baseline. It may take a few column volumes for the new columns to equilibrate.

Solvent exchange and storage of the column should be performed according to the column manufacturer's instructions.

Post-column filter

The Viscotek TDA systems incorporate a post-column filter, which is used to collect any particles shed by the column. Although undetected by most conventional detectors, such as UV and RI, these shed particles will scatter significant amounts of light and thus light scattering signals and baselines will be affected.

Replacement of the filter membrane is recommended:

• If the system is running with an increased pressure.
• If a different buffer is to be used.
• Prior to system cleaning.
• After system cleaning.

The post column filter is situated in line between the columns and the

first detector. To replace it, the system flow should be reduced to 0.1 ml/min or stopped. The DC power to the TDA should be turned off and the module removed from the chassis. The filter housing is then removed from the flow path.

Within the housing of the post column filter are two frits within two O-rings, sandwiching a membrane filter (Figure 2). It is generally recommended to replace the filter, frits and o-rings all together. (If the supply of frits is limited, they can be re-used but their orientation must be maintained to prevent previously collected material escaping into the detectors). Typical filters are PTFE for organic solvents and nylon or cellulose for aqueous buffers and 0.2 μm pore size are used.

Having replaced the filter, the inlet of the filter housing is connected to the flow path. Mobile phase should be allowed to pass right through the housing to ensure that no air enters the system before connecting the filter outlet it to the detectors and replacing the module.

Preventive Cleaning

Systems running a single organic solvent do not usually require regular cleaning, unless contaminants are introduced via sample injection. Aqueous systems require flushing with a cleaning solution about once a month. Cleaning and flushing a system can take up to a few days to achieve maximum benefits and can take place over a weekend, for example. This helps prevent the build up of most contaminants, but will also alleviate symptoms of minor contamination such as elevated baselines and noise or pump backpressure increases.

Cleaning an aqueous system

The cleaning process must be treated in the same way as a solvent changeover. Therefore, before carrying out this procedure, please ensure that Section 4: 'Solvent compatibility and changeover' has been read and fully understood.

To clean your system:

• Remove columns and store according to manufacturer's instructions.
• Reconnect the flow path using a union.
• Replace the post-column filter.
• Replace the running buffer reservoir with one of the cleaning solution specified below and allow it to be flushed through the system overnight at a rate of 0.1 ml/min.

(a) Sodium Azide solution (0.02% w/v). A very effective bactericide.

(b) Methanol: water (10% v/v). A useful cleaning buffer, since most aqueous buffers are miscible with this mixture without precipitation of buffer salts. It is however, very important to check miscibility with the type and concentration of buffer beforehand as precipitation can lead to blockages and leaks.

(c) Acetone: water (10:90 v/v).

• After flushing overnight, purge both of the detectors several times and perform at least 8 washes of the autosampler.

Flushing the system with a few hundred milliliters of cleaning solution should remove any contamination in the majority of cases. Progress can be monitored by collecting data and watching the baseline signals of the detectors. After a reasonable period of cleaning, the detector signals should stabilize (green trace in Fig. 3).

• Prepare fresh running buffer and carry out solvent change; remembering to keep the viscosity detector signals on scale by reducing the flow rate. Or simply exchange solvents at a flow rate of 0.1 ml/min, overnight.
• Purge the RI detector several times, wash the autosampler needle and replace the post-column filter.

During this process, the solvent filter and pump filter may also be replaced although this is not always necessary.

3. System storage

If the system will not be used for an extended period of time - e.g. over a holiday period - then the following protocol for system shutdown should be followed. This will prevent salt deposition, particulate and microbial contamination, as well as preserving solvent.

Aqueous Systems

Using pure water, follow Steps 1-5 in the protocol outlined in 'Cleaning an aqueous system'. Also ensure that the detectors are purged with at least 20 mL of clean solvent and that 8 washes of the autosampler needle are carried out. Repeat these steps with a methanol: water solution (10% v/v). Stop the pump and stopper the inlet and outlet ports located at the front of the TDA chassis. Then switch off the detectors and instrument.

To restart the system, use the protocol outlined in 'Solvent compatibility and changeover'. The system may require cleaning.

Organic Systems

Generally organic systems can be left recirculating at 0.1 ml/min in pure solvent - e.g. THF, DMSO - for up to one month in the same solvent. Beyond this time, the risk of solvent oxidation increases. It is recommended that the user change the solvent and continue recycling until such time that the instrument is used for analysis.

Please contact the local Malvern office for advice about shutdown procedures for organic systems.

4. Solvent compatibility and changeover

It may be necessary to use a different solvent dependent on application or during cleaning. Changing solvents requires care and patience to ensure that the changeover is complete and the system is not damaged by the process.

During solvent changeover, the viscometer bridge and pressure transducers are at significant risk of damage. The nature of the viscometer setup means that when the new solvent reaches the viscometer, a significant signal response occurs. This signal will persist for as long as the delay column takes to fill with the new solvent. Once this has happened, the detector will return to its baseline level. This can be seen in the blue trace in figure 3. When the new mobile phase reaches the detector at 30 ml the signal increases from 170 mV to 350 mV. This returns to 170 mV when the delay column has filled with the new mobile phase at 42.5 ml. The magnitude of the signal shift is dependent on the relative viscosities of the two different solvents.

During solvent changeover, the flow rate must be reduced so that the viscometer signals, both IP (inlet pressure) and DP (differential pressure) remain on scale at all times. (We recommend that the gain is set to 2.5V in the 'Configure…' dialogue so that the full detector scale can be seen.) The simplest way to guarantee a successful solvent changeover is to set the flow rate to 0.1 ml/min and allow the solvent exchange to occur overnight.

Once the viscometer bridge has rebalanced, the flow rate can be accelerated to the running mode flow rate; but, always ensuring that the viscometer signal stays on scale. The refractive index and viscosity detectors should be purged repeatedly to ensure they too are flushed out.

Figure 3: Chromatogram showing detector signals from refractive index (red), viscometer DP (blue) and viscometer IP (dark yellow).

This should be considered every time the solvent is changed, whether for instrument cleaning or to run the chromatography in a different mobile phase.

Solvent compatibility

Viscotek systems are compatible with most solvents and buffers.

Aqueous solvents are solutions prepared in pure water and do not contain any organic solvents or alcohols. Aqueous solvents are largely interchangeable since their viscosities are not significantly different.

If a system running in an aqueous solvent is changed for another solvent containing an organic component then the main concerns are salt precipitation at the buffer interface and damage to the viscometer. To minimize salt precipitation, we recommend that the system is flushed with >50mL of pure water prior to exchange with the new solvent. To prevent damage to the viscometer, follow the recommended protocols outlined in the next subsection 'Solvent changeover: Aqueous to Aqueous' and Steps 2 and 3 in Organic to Aqueous'.

Organic solvent exchange is more complicated due to the vastly different viscosities of organic solvents. Generally, it is recommended to use one system for a given organic solvent. If this is not possible, the exchange of the solvents this should be carried out as outlined in the next subsection 'Solvent changeover: Organic to Organic'.

It is strongly recommend that users check the miscibility of the solvents being interchanged. Common GPC/SEC organic solvents are summarized in Table 3. These solvents are all miscible and may be interchanged according to the described protocol.

Solvent changeover

Aqueous to Aqueous

• Remove analytical columns and replace with a union.
• Replace post-column filter.
• Change the solvent reservoir to pure, filtered water and draw 30mL through the degasser.
• Pump at least 50mL of pure, filtered water - no salt or buffer - through the system at flow rate between 0.1 and 1 mLmin-1. Keep viscometer signals on scale.
• Purge the RI and DP with 20mL of pure water.
• Change the solvent reservoir to the desired solvent and draw 30mL through the degasser.
• Pump the new solvent through the system between 0.1 - 1.0 mLmin-1. Keep viscometer signals on scale.
• Purge the RI and DP with at least 20mL of new solvent.
• Increase the flow rate to the normal running mode and allow the detector baselines to stabilize before injecting standards or samples.

Aqueous to Organic

• Follow Steps 1-5 detailed in the previous subsection 'Aqueous to Aqueous'.
• Change the solvent reservoir to the desired organic solvent e.g. 10% methanol solution, acetone, acetonitrile etc.
• Draw 30mL through the degasser.
• Pump the new solvent through the system at flow rates between 0.1 - 1.0 mLmin-1, keeping the viscometer signals on scale. This can be done overnight at a flow rate of 0.1 mLmin-1.
• Purge both the RI and DP with at least 20mL of the new solvent.
• Increase the flow rate and allow the detector baselines to stabilize before injecting standards or samples.

Organic to Organic

• Remove analytical columns and replace with a union.
• Replace post-column filter.
• Change the solvent reservoir to acetone and draw 30mL through the degasser.
• Pump at least 120mL acetone through the system at flow rate between 0.1 and 1 mLmin-1. Keep viscometer signals on scale. This can be done overnight at a flow rate of 0.1 mLmin-1.

• Purge both the RI and DP with at least 20mL acetone.
• Change the solvent reservoir to the desired solvent and draw 30mL through the degasser.
• Pump the new solvent through the system at flow rates between 0.1 - 1.0 mLmin-1, keeping the viscometer signals on scale. This can be done overnight at a flow rate of 0.1 mLmin-1.
• Purge the RI and DP with at least 20mL of the new solvent.
• Increase the flow rate to the normal running mode and allow the detector baselines to stabilize before injecting standards or samples.

Organic to Aqueous

It is strongly recommended not to switch one system from running organic to aqueous solvents or viceversa. Although possible, this can lead to extended downtime and more frequent issues than when a system is kept solely in organic or aqueous solvents. The only exception to this may be during the installation where an instrument has been shipped in THF and is to be run in aqueous mobile phase. If a system is to be switched from THF to water, then the protocol outlined in 'Solvent changeover: Organic to Organic' must be followed.

5. Common symptoms of rising issues

There are a number of symptoms which indicate when some routine maintenance might be required.

Increasing pump pressure

The most common symptom a system may display is an increase in pump pressure. Over time, one of the filter elements within the system may get blocked.

##Table 3: Commonly used organic solvents and their typical running temperatures. All of these solvents are miscible.

Within a GPCmax, there is a filter before the injection valve. If the pump shows a pressure above ~ 0.4 MPa / 60 psi with just a wide bore waste line attached to the GPCmax outlet at a flow rate of 1 mL/min, then this post-pump filter should be replaced.

Table 3: Common organic solvents used in GPC/SEC.

 The post-column filter will become blocked with particles shed from the column and this will be noticeable in the GPCmax pump pressure. To test this, disconnect and reconnect the filter housing to the flow path. If the pump pressure is above 1 MPa, the filter needs to be replaced. Blockage of the filter can take anywhere from days to months depending on the quality of the column and can also be affected by poor sample or mobile phase quality. If the pressure has increased, the filter should be replaced as described below.

Other issues that may cause an increase in pump pressure can include debris build up in the pump heads and also degraded columns. If none of these improves the situation, it is best to contact a Malvern representative.

Light Scattering Detector Noise

Another common symptom is an increased level of noise in the light scattering detectors. This is caused by particles of debris passing through the system. Since they are generally larger than the molecules of interest in the sample, they cause spikes in the light scattering detectors and the low angle detector (LALS) is always more sensitive to this than the right angle detector (RALS).

Figure 4: Comparison of RALS signals pre- and post-cleaning. prior to cleaning (black) the RALS signal contains many spikes from dirt in the system. Post cleaning (green) the signal is much more stable.

To determine whether or not a system requires cleaning, turn off any digital filtering and de-spiking functions in the software and then monitor the light scattering signal. Software de-spiking and digital filtering algorithms are useful for improving data but can sometimes mask possible issues that, if visible, would indicate a need for system cleaning.

Using a fresh and filtered buffer, collect data for a period of 30 minutes and monitor the baselines of the detector signals.

Figure 4 shows a comparison of the RALS signal for a PBS run buffer pre- and post-cleaning. As evident in this figure, the pre-cleaning signal, in the absence of any de-spiking or digital filtering, shows a large degree of signal spiking across the entire run. If the baseline light scattering signals resemble this, then the system needs some level of cleaning. After cleaning, the signal spiking is infrequent and the results are consistent with those typical of newly installed systems.

Elevated light scattering baselines

As well as becoming noisy, the light scattering detector baselines may themselves increase. For instance, a new system would be expected to have RALS and LALS signals beneath 50 mV when running only solvent and depending on the solvent in use. If the light scattering cell becomes contaminated or dirty, these, and particularly the LALS detector baseline may increase. This will eventually reduce the dynamic range of the instrument and if they increase above the upper limit of the detector, will make it unusable. The light scattering cell itself may then require cleaning.

If this happens it is best to contact your local Malvern representative for advice.

Noise in other detector signals

It is less common for the other detectors to become noisy however it can occur.

Noise or drift in the viscosity detector or the refractive index detector could be due to an air bubble trapped in the system, dripping from the waste line, incomplete solvent exchange or a poorly mixed solvent.

To rectify these, ensure the waste line is not dripping, perform a series of purges of both the RI and the DP, place the mobile phase reservoir on a magnetic stirrer and allow an extended period of equilibration.

Poor chromatography

If the detector baselines are changing or becoming noisy after sample injection, this may be due to the sample itself. If the peak is not a nice shape and shows significant tailing, most notably in the light scattering detectors, or the descending edge of the peak is noisier than the ascending edge, it is likely that the sample is interacting with the column and is not being separated solely by size, which is a pre-requisite for GPC/SEC. In this situation different mobile phase compositions should be used to try to optimize the chromatography to obtain the best results. It may also be worth using a different type of post-column filter as the interaction may be here.

Poor analytical conditions can also lead to partial deposition of sample material within the detector system. This can affect the light scattering cell, most notably by permanently increased scattering background (the baseline). This can also affect the viscometer delay column. Deposits within the delay column lead to a permanent negative bias of the viscometer DP signal.

If the detector signal returns to its usual baseline after an extended run time, the detector itself is not heavily affected. The most likely consequence of sample/column interaction is poor data and a requirement to clean out the system more frequently.